There is no one way to do mixed-methods research. As we noted earlier, mixed-methods research is pragmatic, and the questions influence the method selected. A study may be primarily quantitative with a part that is qualitative. Conversely, it may be primarily qualitative with a quantitative component. Of course, it may be anything in between (see Morse, 2003, for an example of the many possible weightings of qualitative and quantitative research within one study). Figure 20.1 presents a continuum of mixed-methods research. The balance between quantitative and qualitative components will depend on the study, as will the specific methods used within each part of the study.
A mixed-methods study might collect descriptive quantitative measures and qualitative interview data based on an aspect of the quantitative data. An experimental intervention could be combined with a qualitative study evaluating participants’ perception of the intervention. Interviews and focus groups could be supplemented by descriptive data. The permutations of mixed-methods designs are infinite. Any combination of questions that suggest mixed methods can be accommodated so long as the researcher or research team has the skills to conduct all parts of the study.
Although many mixed-method designs are possible (e.g., see Creswell, 2009a, or Leech & Onwuegbuzie, 2009, for discussions of other ways to describe designs) the way in which a study is conducted often is characterized in one of two ways (Creswell, 2009b; Tashakkori & Teddlie, 1998; Teddlie & Tashakkori, 2009). The first, and most common, type of mixed-methods design uses a staggered arrangement of quantitative and qualitative components. In this instance, called sequential mixed methods, one part comes first and the other part of the study follows. Either the quantitative or qualitative part of the study can come first. Often the results of the first part of the study influence what occurs in the second part of the study. For example, a researcher might collect children’s fitness test scores to provide a description of the fitness levels in some group. Based on the BMI scores above 30, participants would be recruited for semistructured interviews focusing on how they perceived that their body composition and fitness levels affected their interactions with their peers.
When the quantitative and qualitative components occur at the same time or are independent, the design commonly is labeled parallel, or concurrent, mixed-methods research. From the beginning of the study both methodological approaches are used, and the results of one part do not dictate the participants or methods for the second part. For example, a research team may be examining the influence of physical activity and diet on various health indicators, such as blood pressure, heart rate, fitness, and blood lipids, with participants engaging in physical activity of various intensities and diet programs that range from easy to implement to those that require large changes in eating behavior. The researchers want to know not only the effectiveness of various intervention combinations but also the perceptions of those who are successful and not successful in maintaining the diet and physical activity interventions to which they were assigned. The qualitative portion of the study would use observation and interviews to address this question. The results could indicate that one combination of diet and physical activity was the most effective in improving the health indicators but that another program was nearly as effective and was perceived by the participants as more enjoyable.